AI Dictionary › AI Fundamentals

Autoencoder

An autoencoder is a neural network that learns to compress data and then reconstruct it, without external labels. It has two parts: an encoder that reduces the input to a compact representation, the latent space, and a decoder that tries to rebuild the original input from it. It's like asking someone to summarize a book in a few sentences and then rewrite its plot from that summary alone: to succeed they must capture the essence, discarding superfluous detail. By training to minimize the difference between input and reconstruction, the model discovers the data's most meaningful features.

Definition

Autoencoders are used for dimensionality reduction, image denoising, and anomaly detection, since an anomalous input reconstructs poorly. Their variants, like variational autoencoders, are also generative and form an important conceptual building block for many modern representation-learning models.

Related terms

• Encoder-Decoder
• Diffusion Model
• Embedding
• Generative Adversarial Network (GAN)
• Deep Learning

More in AI Fundamentals

• LLM (Large Language Model)
• Token
• Context Window
• Temperature
• AI Hallucination
• Fine-tuning
• RAG (Retrieval Augmented Generation)
• Embedding

Put it into practice

• Real scenario: apply this concept with an instant AI verdict.
• Daily gym: short drills to train dictionary concepts.
• Playoff game: challenge other players, elimination-style.

📱 Download the Android app (beta) iOS coming soon

Say what you mean. Get what you need.

Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.